The 451 Group reports on TokuDB v3.0 for MySQL

Posted On March 5, 2010 | By John Partridge | 2 comments

Matt Aslett over at The 451 Group has written a Market Development report entitled “Tokutek delivers ACID transaction support with TokuDB version 3.0.” Get the full report and you can try out The 451 Group’s services for free by visiting http://the451group.com/apply/apply.php.

Matt observes that, “While TokuDB is effectively an operational database technology, it does blur the lines between operations and analytics”. We agree with that perspective and that blurred line is particularly relevant to the community of advanced web application developers who built on MySQL.

Ian Eure over at Digg articulated the issue well: “The fundamental problem is endemic to the relational database mindset, which places the burden of computation on reads rather than writes. This is completely wrong for large-scale web applications, where response time is critical.”

Because of the B-tree index’s inability to perform at high speed simultaneously on read-intensive and write-intensive workloads, database usage has clumped around the two ends of the trade-off: read-intensive OLAP and write-intensive OLTP. Those are big “clumps” indeed and the valley in between – where advanced MySQL based web applications reside – is sizable.

At Tokutek we are serving those applications with Fractal Tree indexing and beginning to fill the valley. Or as Mark Callaghan noted in his MySQL+memcached piece, “Write-optimization has finally arrived for MySQL with the availability of TokuDB.”

2 thoughts

  1. It would be great to learn more about TokuDB at the MySQL Conference. But I don’t see a talk listed for TokuDB.

  2. Bradley C. Kuszmaul says:

    We submitted two proposals.
    1) Michael Bender and I proposed to explain how fractal trees work. This talk is a big revision of the talk I did at the OpenSQL camp. This talk was declined without explanation.
    2) Sheeri Cabral and I proposed to talk about practical issues that one faces when dealing with hundreds of millions or billions of rows. This talk is in Limbo.

    I’m still hoping to present these talks…

Leave a Reply

Your email address will not be published. Required fields are marked *